ObjectiveLevo-α-acetylmethadol (LAAM, levacetylmethadol) is a long-acting opioid agonist used for the prevention of opioid withdrawal. LAAM undergoes sequentialN-demethylation to norLAAM and dinorLAAM, which are more potent and longer-acting than LAAM. Hepatic and intestinal microsomalN-demethylationin vitrois catalysed mainly by cytochrome P450 (CYP) 3A4; however, the role of CYP3A in LAAM disposition in humansin vivois unknown. This investigation tested the hypothesis that CYP3A induction (or inhibition) would increase (or decrease) LAAM metabolism and bioactivation and, thus, clinical effects. It also related changes in LAAM disposition during enzyme inhibition or induction to any changes in pharmacological effect.MethodsHealthy volunteers (n = 13) completed the three-way, randomised, balanced crossover study. Subjects received oral LAAM (0.25 mg/kg) after CYP3A induction (rifampicin [rifampin]), inhibition (troleandomycin) or nothing (controls). Plasma and urine LAAM, norLAAM and dinorLAAM were determined by electrospray high-performance liquid chromatography/mass spectrometry (HPLC/MS). Dark-adapted pupil diameter change from baseline (miosis) was the LAAM effect measure. Results were analysed by noncompartmental methods and by a combined pharmacokinetic/pharmacodynamic model.ResultsCompared with controls, CYP3A induction (or inhibition) decreased (or increased) plasma LAAM concentrations and mean area under the plasma concentration-time curve from time zero to infinity (AUC∞199 ± 91 [control] versus 11.3 ± 4.0 [rifampicin] and 731 ± 229 ng · h/mL [troleandomycin]; p < 0.05), and increased (or decreased) median formation clearances of norLAAM (1740 versus 14 100 and 302 mL/h/kg; p < 0.05) and dinorLAAM (636 versus 7840 and 173 mL/h/kg; p < 0.05). Surprisingly, however, CYP3A induction (or inhibition) decreased (or increased) mean plasma metabolite AUC from 0 to 96 hours (AUC96) [norLAAM + dinorLAAM] (859 ± 241 versus 107 ± 48 and 1185 ± 179 ng · h/mL; p < 0.05) and clinical effects (mean miosis AUC96128 ± 40 versus 22.5 ± 14.9 and 178 ± 81 mm · h; p < 0.05). Clinical effects were best correlated with plasma norLAAM concentrations.ConclusionCYP3A mediates human LAAMN-demethylation and bioactivation to norLAAM and dinorLAAMin vivo. Paradoxically, however, CYP3A induction decreased and inhibition increased LAAM active metabolite concentrations and clinical effects. This suggests a CYP3A-mediated metabolic pathway leading to inactive metabolites, which predominates over CYP3A-dependent bioactivation. These results highlight the need for clinical investigations to validatein vitrodrug metabolism studies.